27 Twitter 1431050 2016 3 14
1 Twitter,,.,.,., Twitter,.,,.,,. URL,,,. BoW(Bag of Words), LSI(Latent Semantic Indexing)., URL,,,,., Accuracy, AUC(Area Under the Curve), Precision, Recall, F,. URL,,,., 3.,,,.
2 1 7 1.1........................................... 7 1.2.......................................... 7 1.3.......................................... 8 1.4......................................... 9 1.5...................................... 9 2 10 2.1 [1]............................ 10 2.1.1................................ 10 2.1.2............................ 10 2.1.3................................. 11 2.2 Twitter [2].......................... 11 2.2.1................................. 11 2.3 Twitter..................... 12 2.3.1 [3]..................... 12 2.3.2 [4]......................... 13 2.3.3................................. 13 2.4 Twitter........................ 13 2.4.1 [5].................. 13 2.4.2................................. 14 2.4.3 [6]....................... 14 2.5............... 15 3 16 3.1........................................... 16 3.2.................................. 16 3.3.......................... 16 3.4....................................... 17 3.5............................... 17 3.6......................................... 18 3.6.1 URL................................... 19 3.6.2................................. 19 3.6.3................................. 19 3.7.................................... 19 3.7.1.............................. 19
3 3.7.2 K................................. 20 3.8.................................... 20 3.8.1 Accuracy( )................................ 22 3.8.2 Precision( )................................ 22 3.8.3 Recall( )................................. 22 3.8.4 F1-score(F )................................. 23 3.8.5 AUC(Area Under the Curve)......................... 23 4 24 4.1...................................... 24 4.2........................................... 24 4.2.1........................ 24 4.2.2.............................. 26 4.3........................................... 26 5 27 5.1......................................... 27 5.2...................................... 27 5.2.1.................................. 27 5.2.2............................. 27 5.2.3..................................... 30 6 31 6.1.............................. 31 6.2........................... 31 6.3 LSI................................. 32 6.4......................................... 34 6.4.1 URL................................... 34 6.4.2................................. 34 6.4.3................................. 34 6.4.4................................. 34 7 41 7.1 LSI............................ 41 7.2......................................... 41 7.2.1 URL................................... 41 7.2.2................................. 41 7.2.3................................. 42 7.3................................. 42 7.3.1.................... 42 7.3.2.................... 43 7.3.3 [2].............................. 44
4 8 45 8.1.......................................... 45 8.2....................................... 45 8.2.1................................. 45 8.2.2........................... 45 8.2.3............................. 45 8.2.4............................. 46 8.2.5............................. 46 47 49 49
5 1.1....................... 8 2.1....................... 12 2.2 20........................ 14 3.1.................................... 17 3.2.............................. 20 3.3.................................... 21 3.4.............................. 21 3.5......................... 22 3.6 ROC AUC.................................... 23 4.1 ROC......................... 25 4.2 ROC......................... 25 4.3 ROC.......................... 25 5.1.................................. 29 6.1 ( )........................... 39 6.2 (+URL)................................. 39 6.3 (+ )................................ 40 6.4 (+ )........................... 40 6.5 (+all)................................... 40
6 2.1 [2]................................... 12 3.1............................ 18 3.2.................................... 20 4.1 1......................... 24 4.2 2......................... 26 4.3................................ 26 5.1................................ 28 6.1............................... 32 6.2 1( ).................... 33 6.3 2( ).................... 33 6.4 ( )........................... 33 6.5 (+URL)........................... 35 6.6 2(+URL).......................... 35 6.7 (+URL)................................. 35 6.8 1(+ )......................... 36 6.9 2(+ )......................... 36 6.10 (+ )................................ 36 6.11 1(+ )........................... 37 6.12 2(+ )........................... 37 6.13 (+ )................................. 37 6.14 1(+all)............................ 38 6.15 2(+all)............................ 38 6.16 (+all)................................... 38 7.1.......................... 42 7.2........ 43 7.3.......................... 43 8.1 (True Positive).......................... 49 8.2 (True Negative)......................... 50 8.3 (False Positive).......................... 51 8.4 (False Negative)......................... 52
7 1 1.1 SNS., SNS. Twitter 1,,., Twitter,.,.,.,,. Twitter,. Twitter,. 1.2, 1.,.,,.,.., Twitter,., Twitter,.. Twitter,. ( 1.1).,,..,. 1 https://twitter.com/
1 8,,. 1.1: 1.3,..,,., 3 URL,,.
1 9 1.4.,,,,.,.,,,.,,., Twitter,,, ( ),. 1.5. 1,,,,. 2,. 3,. 4, 2 2. 5,. 6,. 7,. 8,.
10 2, Twitter. 2.1 [1] 2.1.1 [1],.,., ( ).,,.. 2.1.2,. 4...,..,..,,,. ( ),.,.
2 11 2.1.3,.,,.. 2.2 Twitter [2] Sungho Jeon [2], Twitter 4, SVM,., 12 176426., 5618..,... URL URL URL. URL URL. URL. URL,.,..,,,.,.,.,. 2.1., URL F.. 2.2.1 SVM.,,. 4 SVM
2 12 2.1: [2] Step Entered feature Recall Precision F-score 1 0.3084 0.8291 0.4496 2 + 0.5653 0.8611 0.6826 3 +URL 0.7699 0.7957 0.7825 4 + 0.7697 0.7987 0.7839, F,.,,. 2.3 Twitter 2.3.1 [3] [3],,. 2.1. B 2, B. A,. 2.1: ( ).,.,., ( ) ( ), ( ) ( ).,,,.
2 13.,.,,. 2.3.2 [4] [4], ( ).,. A, A ( ). A, A.,,..,.. 2.3.3.,.,.,. 2.4 Twitter 2.4.1 [5] [5],, TV. 2.2 20 TV.,. 2.2.,,. Twitter., SVM,. 3.,.,,,, 5.,.
2 14 2.2: 20 1,.,.,.,,. 2.4.2, 1.,..,.,,. 2.4.3 [6] [6], 2,..
2 15,.,, 2-1.,,.,.,.,..,.,,,.,,.,.,.,. 2.5 Samuel Brody [7],.,,,.,,.,.
16 3 3.1 (SVM),...,, Bag of Words.,,.,. SVM. K (K-fold cross-validation). 3.2, twitter. twitter 1. 3.1., (OR NOT,,, ), ( ),, (, )., 2 /twport 3 Web. 3.3 2.,.,, (,, ),,,.,,,,,,. 3.1. 1, 0. 1 https://twitter.com/search-advanced 2 http://hashtagcloud.net/ 3 http://twport.com/
3 17 3.1: 3.4, MeCab 4..,,. 3.5 Bag of Words 5.,., Latent Semantic Indexing(LSI) 6 7. 1 128. 4 http://mecab.googlecode.com/svn/trunk/mecab/doc/index.html 5 5.2.1 2100 BoW. 6 Latent Semantic Analysis(LSA). 7 http://lsa.colorado.edu/
3 18 3.1: #joqr #npb #allstar 1 HR! #allstar 1 1 http://t.co/xaxqyppc59 #npb 2 1 #allstar #seibulions #npb #AllStarGame 1 ( )! #allstar 1 #npballstar #allstar #npb #tvasahi 1 MVP #carp #npb #npballstar # 1 #joqr #allstar #npb 1 1 http://t.co/3apjtral57 #npb #npballstar #allstar #hanshin #tigers 0 #joqr #npb #allstar 0 #joqr #npb #allstar 0 http://t.co/i9thnmfnem 0 #npb http://t.co/skkuza0hdl #AllStarGame #AllStar 0 #NPB #npballstar #npb 0 #allstar #npb 0 #allstar 0 #allstar 0 #npb 0 #NpbALLSTAR LSI, 2., LSI 2 2. 2 4. 3.6.
3 19 3.6.1 URL 2.2 [2], URL,., Twitter, URL.. URL,. 3.6.2 2.5 [7],.,.,.,. 4.. 3.6.3.., 10. = ( ) 3.2.,,,.,., 3.2,. Yuxin Peng [8], SVM 2,.,,. 3 3.3. 3.7 K. 3.7.1. 1,,. 3.4. 7 3,.
3 20 3.2: 3.7.2 K K, K, K-1, 1. K, K-1 1, K.,,. 3.5., 10,. 3.8, Accuracy( ), Precision( ), Recall( ), F1-score(F ), Area Under the Curve(AUC). 3.2: True Positive(TP) False Positive(FP) False Negative(FN) True Negative(TN)
3 21 3.3: 3.4:
3 22 3.5: 3.8.1 Accuracy( ) Accuracy,.. Accuracy = T P + T N T P + T N + F P + F N 3.8.2 Precision( ) Precision,,. Web,, Web Precision. Recall, F.. P recision = T P T P + F P 3.8.3 Recall( ) Recall,,. Web, Web, Web Recall. Precision, F..
3 23 Recall = T P T P + F N 3.8.4 F1-score(F ) F,, Precision Recall.. F 1 score = 2Recall P recision Recall + P recision = 2 T P T P +F N T P T P +F P = T P T P +F N + T P T P +F P 2T P 2T P + F P + F N 3.8.5 AUC(Area Under the Curve) Area Under the Curve,, ROC(Receiver Operating Characteristic). ROC, True Positive Rate, False Positeve Rate,,. AUC. 3.6 ROC AUC. ROC AUC. 2, AUC 3.6., 3.6 3.6. ROC AUC. AUC. 3.6: ROC AUC
24 4 LSI 2, 2. 4.1, (2014/11/14 ), (2015/7/3 ), (2016/1/2 ) 3000.,, 3, 2. 7:3 10, 2. 4.2 4.2.1,, 3 4.1, 4.2., ROC 4.1, 4.2, 4.3. 4.1: 1 Precision Recall F1-score class baseball 0.95 1.00 0.97 class wimbledon 1.00 0.95 0.97 average 0.97 0.97 0.97 Precision Recall F1-score class wimbledon 0.98 1.00 0.99 class hakone 1.00 0.98 0.99 average 0.99 0.99 0.99 Precision Recall F1-score class baseball 0.98 1.00 0.99 class hakone 1.00 0.98 0.99 average 0.99 0.99 0.99
4 25 4.1: ROC 4.2: ROC 4.3: ROC
4 26 4.2: 2 Accuracy 0.971 0.989 0.989 AUC 0.996 0.999 0.999 4.2.2, 3 10 4.3. 4.3: Accuracy 0.973(+/- 0.012) 0.990(+/- 0.008) 0.990(+/- 0.009) AUC 0.996(+/- 0.003) 0.999(+/- 0.001) 0.998(+/- 0.002) Precision 0.974(+/- 0.011) 0.990(+/- 0.008) 0.990(+/- 0.008) Recall 0.973(+/- 0.012) 0.990(+/- 0.008) 0.990(+/- 0.009) F1-score 0.973(+/- 0.012) 0.990(+/- 0.008) 0.990(+/- 0.009) 4.3, LSI 2, 2. 2.
27 5 5.1,,. Python2.7. MeCab. (BoW, LSI) Python gensim 1. Python scikit-learn 2. 5.2 5.2.1 2015 7 2 Twitter, 5500,..,,,,.,,,,,,., 1:10. F (3.6.3 ).,, 10. 5.2.2, SVM,.,... URL,, #. URL URL. http:// https://,. 1 https://radimrehurek.com/gensim/ 2 http://scikit-learn.org/stable/
5 28,, @. @,.,, ( ) RT @ ( ). RT 2,., #. # 1.. MeCab...,,,... [5].,,,,.. 5.1. 5.1:, 5.1.,...,,.,.,.,..
5 29 5.1:, 6.2.. Bag of Words (BoW ),. Bag of Words Bag of Words Python gensim., N. BoW. 3...,,,,,,,,, 3. 1, 1, 2, 1, 0, 0, 0, 0, 0, 0, 0 0, 1, 0, 0, 1, 1, 1, 1, 1, 0, 0 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1 BoW. 11, 11 BoW. BoW. 3.5 LSI, 1 128.
5 30 3.6 URL,,. LSI 3. 5.2.3, 3.1.,,. ( ), ( ).,,.
31 6 6.1,, LSI.,,,. 2100., 2100,.,. 6.2., %., 5.2.2., LSI 2, 3.6. 0 5% 0.5%, 75 100% 5%. 5 Accuracy F. Accuracy 0.570 0.575, F 0.543 0.550. F, 10 5 6.1. 6.1, Accuracy F.. 1.0% 80%, Accuracy 1, F 3,. 1.0%, 80%.
6 32 6.1: (%) (%) Accuracy Accuracy F F 1.5 95 0.572 0.011 0.545 0.006 3.0 100 0.571 0.007 0.544 0.007 1.0 80 0.571 0.005 0.545 0.008 5.0 95 0.570 0.010 0.544 0.008 2.5 85 0.573 0.013 0.548 0.008 0.5 75 0.572 0.005 0.547 0.010 3.5 90 0.571 0.007 0.545 0.010 3.5 80 0.571 0.010 0.545 0.010 0.5 100 0.573 0.011 0.548 0.010 3.5 85 0.572 0.012 0.546 0.010 3.0 85 0.571 0.032 0.544 0.033 2.5 90 0.573 0.029 0.548 0.034 0.0 95 0.575 0.036 0.550 0.034 4.0 80 0.571 0.037 0.545 0.036 2.0 95 0.572 0.038 0.546 0.041 6.3 LSI LSI 1 128,.,..,. 6.2 6.3, 6.4 6.1., Accuracy 32. 32, 32. 32 Accuracy 32. AUC,. Accuracy 32. F, 32,. 32, F 32., 32.
6 33 6.2: 1( ) 1 precision recall f1-score 16 precision recall f1-score 0.53 0.82 0.64 0.54 0.85 0.66 0.65 0.31 0.42 0.71 0.33 0.45 avg/total 0.59 0.56 0.53 avg/total 0.62 0.58 0.55 2 precision recall f1-score 32 precision recall f1-score 0.55 0.82 0.66 0.60 0.83 0.69 0.66 0.34 0.45 0.73 0.46 0.56 avg/total 0.61 0.58 0.56 avg/total 0.66 0.64 0.63 4 precision recall f1-score 64 precision recall f1-score 0.57 0.81 0.6 0.58 0.93 0.72 0.66 0.37 0.48 0.83 0.35 0.50 avg/total 0.61 0.59 0.57 avg/total 0.71 0.64 0.60 8 precision recall f1-score 128 precision recall f1-score 0.56 0.83 0.67 0.58 0.93 0.71 0.67 0.35 0.46 0.81 0.32 0.46 avg/total 0.62 0.59 0.57 avg/total 0.69 0.62 0.58 6.3: 2( ) 1 2 4 8 Accuracy 0.559 0.584 0.592 0.591 AUC 0.569 0.586 0.593 0.611 16 32 64 128 Accuracy 0.577 0.641 0.637 0.622 AUC 0.644 0.685 0.710 0.723 6.4: ( ) Accuracy AUC Precision Recall f1score 1 0.571 (+/- 0.023) 0.593 (+/- 0.035) 0.596 (+/- 0.029) 0.572 (+/- 0.023) 0.543 (+/- 0.024) 2 0.572 (+/- 0.030) 0.573 (+/- 0.039) 0.595 (+/- 0.031) 0.574 (+/- 0.027) 0.547 (+/- 0.036) 4 0.589 (+/- 0.034) 0.581 (+/- 0.048) 0.609 (+/- 0.038) 0.591 (+/- 0.032) 0.571 (+/- 0.036) 8 0.600 (+/- 0.038) 0.610 (+/- 0.052) 0.622 (+/- 0.040) 0.601 (+/- 0.036) 0.583 (+/- 0.042) 16 0.595 (+/- 0.039) 0.649 (+/- 0.043) 0.627 (+/- 0.050) 0.597 (+/- 0.039) 0.570 (+/- 0.043) 32 0.639 (+/- 0.021) 0.688 (+/- 0.029) 0.661 (+/- 0.027) 0.640 (+/- 0.026) 0.626 (+/- 0.025) 64 0.633 (+/- 0.029) 0.706 (+/- 0.044) 0.701 (+/- 0.037) 0.635 (+/- 0.026) 0.601 (+/- 0.031) 128 0.631 (+/- 0.027) 0.726 (+/- 0.031) 0.698 (+/- 0.013) 0.633 (+/- 0.018) 0.598 (+/- 0.031)
6 34 6.4 6.4.1 URL 1 128 URL.,. 6.5 6.6, 6.7 6.2., 6.3., 6.3, 32. 6.4.2 1 128.,. 6.8 6.9, 6.10 6.3., 6.3 1 16., Accuracy F 8, AUC 32, 6.3 6.4.1.. 6.4.3 1 128.,. 6.11 6.12, 6.13 6.4., 6.3., 6.4.2 1 16., Accuracy, AUC, F 32, 6.3 6.4.1 32. 6.4.4 1 128.,. 6.14 6.15, 6.16 6.5., 6.3..,. 2, 4.
6 35 6.5: (+URL) 1 precision recall f1-score 16 precision recall f1-score 0.57 0.80 0.67 0.56 0.79 0.66 0.68 0.41 0.51 0.68 0.41 0.51 avg/total 0.62 0.60 0.59 avg/total 0.62 0.60 0.58 2 precision recall f1-score 32 precision recall f1-score 0.56 0.84 0.67 0.61 0.81 0.70 0.69 0.36 0.47 0.72 0.48 0.58 avg/total 0.63 0.60 0.57 avg/total 0.66 0.65 0.64 4 precision recall f1-score 64 precision recall f1-score 0.58 0.78 0.67 0.60 0.90 0.72 0.68 0.45 0.54 0.79 0.39 0.53 avg/total 0.63 0.62 0.61 avg/total 0.70 0.65 0.62 8 precision recall f1-score 128 precision recall f1-score 0.58 0.81 0.67 0.61 0.87 0.72 0.68 0.41 0.51 0.74 0.41 0.53 avg/total 0.63 0.61 0.59 avg/total 0.68 0.65 0.63 6.6: 2(+URL) 1 2 4 8 Accuracy 0.602 0.597 0.616 0.608 AUC 0.627 0.583 0.624 0.634 16 32 64 128 Accuracy 0.597 0.646 0.646 0.647 AUC 0.644 0.698 0.719 0.728 6.7: (+URL) Accuracy AUC Precision Recall f1score 1 0.605 (+/- 0.035) 0.628 (+/- 0.043) 0.626 (+/- 0.034) 0.606 (+/- 0.029) 0.588 (+/- 0.033) 2 0.602 (+/- 0.020) 0.590 (+/- 0.016) 0.630 (+/- 0.015) 0.603 (+/- 0.013) 0.581 (+/- 0.017) 4 0.618 (+/- 0.023) 0.633 (+/- 0.023) 0.638 (+/- 0.022) 0.619 (+/- 0.019) 0.604 (+/- 0.022) 8 0.621 (+/- 0.030) 0.640 (+/- 0.030) 0.638 (+/- 0.032) 0.622 (+/- 0.028) 0.610 (+/- 0.030) 16 0.620 (+/- 0.037) 0.663 (+/- 0.038) 0.644 (+/- 0.037) 0.621 (+/- 0.027) 0.605 (+/- 0.031) 32 0.652 (+/- 0.025) 0.697 (+/- 0.025) 0.668 (+/- 0.031) 0.653 (+/- 0.026) 0.645 (+/- 0.025) 64 0.649 (+/- 0.028) 0.715 (+/- 0.034) 0.695 (+/- 0.033) 0.651 (+/- 0.027) 0.628 (+/- 0.030) 128 0.646 (+/- 0.025) 0.733 (+/- 0.026) 0.694 (+/- 0.023) 0.648 (+/- 0.015) 0.624 (+/- 0.022)
6 36 6.8: 1(+ ) 1 precision recall f1-score 16 precision recall f1-score 0.65 0.78 0.71 0.65 0.78 0.71 0.73 0.58 0.65 0.73 0.58 0.65 avg/total 0.69 0.68 0.68 avg/total 0.69 0.68 0.68 2 precision recall f1-score 32 precision recall f1-score 0.64 0.71 0.67 0.62 0.75 0.68 0.68 0.61 0.64 0.69 0.54 0.61 avg/total 0.66 0.65 0.65 avg/total 0.65 0.65 0.64 4 precision recall f1-score 64 precision recall f1-score 0.67 0.74 0.70 0.59 0.92 0.71 0.71 0.63 0.67 0.83 0.38 0.52 avg/total 0.69 0.69 0.69 avg/total 0.71 0.64 0.62 8 precision recall f1-score 128 precision recall f1-score 0.66 0.75 0.70 0.62 0.86 0.72 0.72 0.61 0.66 0.77 0.46 0.57 avg/total 0.69 0.68 0.68 avg/total 0.69 0.66 0.65 6.9: 2(+ ) 1 2 4 8 Accuracy 0.682 0.655 0.686 0.683 AUC 0.732 0.693 0.710 0.701 16 32 64 128 Accuracy 0.682 0.647 0.643 0.663 AUC 0.732 0.723 0.727 0.726 6.10: (+ ) Accuracy AUC Precision Recall f1score 1 0.676 (+/- 0.029) 0.728 (+/- 0.023) 0.684 (+/- 0.029) 0.677 (+/- 0.030) 0.673 (+/- 0.031) 2 0.665 (+/- 0.034) 0.701 (+/- 0.029) 0.666 (+/- 0.035) 0.665 (+/- 0.034) 0.664 (+/- 0.034) 4 0.684 (+/- 0.030) 0.711 (+/- 0.043) 0.687 (+/- 0.030) 0.684 (+/- 0.030) 0.683 (+/- 0.030) 8 0.687 (+/- 0.027) 0.714 (+/- 0.038) 0.689 (+/- 0.029) 0.687 (+/- 0.028) 0.686 (+/- 0.027) 16 0.676 (+/- 0.029) 0.728 (+/- 0.023) 0.684 (+/- 0.029) 0.677 (+/- 0.030) 0.673 (+/- 0.031) 32 0.664 (+/- 0.026) 0.740 (+/- 0.032) 0.680 (+/- 0.032) 0.665 (+/- 0.027) 0.656 (+/- 0.025) 64 0.660 (+/- 0.019) 0.726 (+/- 0.011) 0.688 (+/- 0.023) 0.662 (+/- 0.018) 0.648 (+/- 0.021) 128 0.644 (+/- 0.033) 0.723 (+/- 0.028) 0.684 (+/- 0.030) 0.646 (+/- 0.027) 0.625 (+/- 0.036)
6 37 6.11: 1(+ ) 1 precision recall f1-score 16 precision recall f1-score 0.70 0.53 0.60 0.70 0.53 0.60 0.61 0.76 0.68 0.61 0.76 0.68 avg/total 0.66 0.65 0.64 avg/total 0.66 0.65 0.64 2 precision recall f1-score 32 precision recall f1-score 0.65 0.65 0.65 0.71 0.59 0.64 0.66 0.66 0.66 0.65 0.76 0.70 avg/total 0.66 0.66 0.66 avg/total 0.68 0.68 0.67 4 precision recall f1-score 64 precision recall f1-score 0.65 0.65 0.65 0.69 0.67 0.68 0.67 0.67 0.67 0.68 0.70 0.69 avg/total 0.66 0.66 0.66 avg/total 0.69 0.69 0.69 8 precision recall f1-score 128 precision recall f1-score 0.68 0.63 0.65 0.65 0.60 0.62 0.67 0.72 0.69 0.63 0.68 0.65 avg/total 0.67 0.67 0.67 avg/total 0.64 0.64 0.64 6.12: 2(+ ) 1 2 4 8 Accuracy 0.647 0.656 0.660 0.675 AUC 0.696 0.713 0.715 0.728 16 32 64 128 Accuracy 0.665 0.677 0.686 0.637 AUC 0.741 0.745 0.762 0.720 6.13: (+ ) Accuracy AUC Precision Recall f1score 1 0.642 (+/- 0.028) 0.692 (+/- 0.023) 0.648 (+/- 0.026) 0.642 (+/- 0.026) 0.638 (+/- 0.028) 2 0.651 (+/- 0.033) 0.702 (+/- 0.036) 0.651 (+/- 0.034) 0.651 (+/- 0.034) 0.651 (+/- 0.034) 4 0.668 (+/- 0.034) 0.723 (+/- 0.046) 0.668 (+/- 0.034) 0.668 (+/- 0.033) 0.668 (+/- 0.034) 8 0.675 (+/- 0.030) 0.731 (+/- 0.031) 0.676 (+/- 0.030) 0.675 (+/- 0.029) 0.674 (+/- 0.029) 16 0.669 (+/- 0.022) 0.731 (+/- 0.023) 0.673 (+/- 0.022) 0.668 (+/- 0.021) 0.666 (+/- 0.021) 32 0.679 (+/- 0.033) 0.751 (+/- 0.037) 0.684 (+/- 0.032) 0.678 (+/- 0.032) 0.676 (+/- 0.034) 64 0.673 (+/- 0.032) 0.751 (+/- 0.038) 0.674 (+/- 0.032) 0.674 (+/- 0.032) 0.673 (+/- 0.032) 128 0.653 (+/- 0.024) 0.727 (+/- 0.021) 0.653 (+/- 0.024) 0.653 (+/- 0.024) 0.652 (+/- 0.024)
6 38 6.14: 1(+all) 1 precision recall f1-score 16 precision recall f1-score 0.76 0.63 0.69 0.75 0.66 0.70 0.68 0.80 0.74 0.69 0.78 0.74 avg/total 0.72 0.72 0.71 avg/total 0.72 0.72 0.72 2 precision recall f1-score 32 precision recall f1-score 0.75 0.70 0.73 0.72 0.67 0.69 0.72 0.77 0.74 0.69 0.74 0.72 avg/total 0.74 0.74 0.74 avg/total 0.71 0.71 0.71 4 precision recall f1-score 64 precision recall f1-score 0.76 0.71 0.73 0.67 0.70 0.68 0.73 0.78 0.75 0.69 0.66 0.67 avg/total 0.74 0.74 0.74 avg/total 0.68 0.68 0.68 8 precision recall f1-score 128 precision recall f1-score 0.74 0.68 0.71 0.66 0.67 0.67 0.72 0.77 0.74 0.66 0.66 0.66 avg/total 0.73 0.73 0.73 avg/total 0.66 0.66 0.66 6.15: 2(+all) 1 2 4 8 Accuracy 0.715 0.735 0.744 0.728 AUC 0.792 0.803 0.816 0.801 16 32 64 128 Accuracy 0.719 0.706 0.678 0.662 AUC 0.789 0.792 0.768 0.744 6.16: (+all) Accuracy AUC Precision Recall f1score 1 0.723 (+/- 0.030) 0.793 (+/- 0.028) 0.728 (+/- 0.030) 0.723 (+/- 0.030) 0.722 (+/- 0.030) 2 0.738 (+/- 0.028) 0.809 (+/- 0.028) 0.739 (+/- 0.028) 0.738 (+/- 0.029) 0.738 (+/- 0.029) 4 0.742 (+/- 0.025) 0.813 (+/- 0.030) 0.743 (+/- 0.027) 0.742 (+/- 0.026) 0.742 (+/- 0.025) 8 0.736 (+/- 0.035) 0.809 (+/- 0.027) 0.738 (+/- 0.034) 0.736 (+/- 0.036) 0.735 (+/- 0.036) 16 0.718 (+/- 0.040) 0.797 (+/- 0.032) 0.724 (+/- 0.038) 0.718 (+/- 0.039) 0.716 (+/- 0.040) 32 0.705 (+/- 0.030) 0.789 (+/- 0.033) 0.706 (+/- 0.029) 0.705 (+/- 0.029) 0.704 (+/- 0.030) 64 0.692 (+/- 0.025) 0.775 (+/- 0.016) 0.692 (+/- 0.025) 0.692 (+/- 0.026) 0.692 (+/- 0.025) 128 0.673 (+/- 0.023) 0.747 (+/- 0.025) 0.673 (+/- 0.024) 0.674 (+/- 0.023) 0.673 (+/- 0.024)
6 39 6.1: ( ) 6.2: (+URL)
6 40 6.3: (+ ) 6.4: (+ ) 6.5: (+all)
41 7 7.1 LSI BoW,. LSI. 1 128,... 7.2 7.2.1 URL URL 1 128. 2.2 URL. 2.2 URL,.,,.,.,.,. URL. 7.2.2,,..., ( ) ( ).,.
7 42 7.2.3,,,..,,...,.,,., ( )( ) ( )( )( ),.,.,. 7.3 7.3.1. 7.1. 7.1, 7.2, 7.3, 1, 0. 0-0, 0 (1 ) 0,.,,.,,.,. 7.1: #joqr #npb #allstar 1 0 18:45 0 1 0 #allstar #npb 1 0 #tvasahi #allstar 1 0 #joqr #npb #allstar 1 1 1 #allstar #npb 1 1. 7.2.,
7 43,. 7.2: #allstar 1 0 #npb #baystars 1 0 #allstar 1 0 5 #baystars #allstar 1 1 #carp #npb 1 1 #AllStar #NPB #npballstar #AllStarGame 1 1 7.3.2,,,.,,,,,. 7.3.,.,.,,,. 7.3: #npb 0 1 #allstar #npb 0 1 #allstar 0 1 #allstar #npb 0 1 #allstar 0 1 #AllStarGame #npb 0 1 #allstar 0 1 #npb #AllStarGame 0 1 #npb #allstar MVP #npb 0 1 0 1
7 44 7.3.3 [2] 2.1 F 6.16 4 F, 0.04., F 0.7,.,.,,.,.,,.,,.,,..
45 8 8.1,.,,. URL,,,. BoW, LSI., URL,,,,., Accuracy, AUC, Precision, Recall, F,. URL,,,., 3.,,,. 8.2 8.2.1 7.3.1.., 0-0, 0 (1 ) 0 0 0. 8.2.2 7.3.2.,,,. 8.2.3, URL,,..,
8 46 Like.. 8.2.4,.,. 8.2.5, SVM.,,.,., URL,,,.
47,,.,,,.., 2,, 1,,,, 4,,,,,..
48 [1].. WISS2010, 41-46, 2010. [2] Sungho Jeon, Sungchul Kim, and Hwanjo Yu. Don t Be Spoiled by Your Friends: Spoiler Detection in TV Program Tweets. Seventh International AAAI Conference on Weblogs and Social Media, 2013. [3],,,. Twitter. 19, 138-141, 2013. [4],,,,,. Twitter.. MVE, 110(457), 165-169, 2011. [5],.. 96 (GN), 2015. [6],. SNS. 96 (GN), 2015. [7] Samuel Brody, Nicholas Diakopoulos. Cooooooooooooooollllllllllllll!!!!!!!!!!!!!!: using word lengthening to detect sentiment in microblogs. Proceedings of the 2011 Conference on Empirical Methods in Natural Language Processing, 562-570, 2011. [8] Yuxin Peng, Jia Yao. AdaOUBoost: adaptive over-sampling and under-sampling to boost the concept learning in large scale imbalanced data sets. Proceedings of the international conference on Multimedia information retrieval. ACM, 2010.
8 49, (True Positive, True Negative, False Positive, False Negative, 25 ). 8.1: (True Positive) 2 #allstar #npb 1 1 1 #allstar #npb 1 1 #NPB #npballstar 1 1 ( ) #NPB #npballstar 1 1 #npb #allstar #npballstar 1 1 #joqr #npb #allstar 1 1 #allstar 1 1 #joqr #npb #allstar 1 1 #npb 1 1 #npballstar #joqr #npb #allstar 1 1 10 1 1 #TVasahi # # #allstar #allstargame #npballstar #npb #allstar 1 1 #npb #allstar #npballstar 1 1 #allstar #npb 1 1 8-3 2015 #allstar #npb 1 1 1 1 http://t.co/lekbpn2q1a #seibulions #npb 3!!2015 1 1 #allstar #TVasahi MVP #carp #npb #npballstar # 1 1 MVP #allstar 1 1 #joqr #allstar #npb 1 1 http://t.co/gsmmymtayc #npb 1 1 http://t.co/mvpabqloxh http://t.co/kyon77l8rk 1 1 #npb 38 http://t.co/sacfnamhft # 1 1 #NPB 1 1 http://t.co/ilaqcd2xyt #npb http://t.co/7bbiidph7a # #NPB 1 1
8 50 8.2: (True Negative)...... #allstar #npb 0 0 #baystars #allstar 0 0 #TVasahi 0 0 # # #allstar #allstargame #npballstar #npb ( ) ( 0 0 )!!!!! #allstar MLB 0 0 http://t.co/1gojndmdda #npb # #npb #AllStarGame 0 0 0 0 #npballstar #allstar #npb!! #allstar 0 0 #BSasahi 0 0 #TVasahi # # #allstar #allstargame #npballstar #npb 2 #TVasahi # # 0 0 #allstar #allstargame #npballstar #npb 0 0 #allstar #AllStarGame #AllStar #NPB 0 0 #npballstar #allstar 0 0 #npb 0 0 #npb 0 0 ww #allstar 0 0 2005 #AllStar #NPB #npballstar 0 0 #AllStarGame AS.520 #npb 0 0 #npb 0 0 http://t.co/2fu06yhtjn 0 0 #npb #npb #allstar 0 0 #allstar 0 0 w #npb 0 0 #allstar 0 0 #TVasahi # # #allstar 0 0 #allstargame #npballstar #npb
8 51 8.3: (False Positive) #allstar 0 1 #npb #allstar 0 1 #npb #AllStarGame 0 1 #npb 0 1 #allstar #npb 0 1 0 1 #npb #joqr #npb #allstar 0 1 #joqr #npb #allstar 0 1 #allstar 0 1 P #allstar #npb 0 1 #allstar 0 1 PL #allstar 0 1 k #allstar #npb 0 1 #allstar 0 1 #joqr #npb 0 1 #allstar #TVasahi # # 0 1 #allstar #allstargame #npballstar #npb #AllStar #NPB #npballstar #AllStarGame 0 1 #allstar 0 1 #allstar 0 1 #allstar #joqr 0 1 #TVasahi # # #allstar #allstargame 0 1 #npballstar #npb 9 #allstar #npb 0 1! #allstar 0 1 20 #npballstar #allstar #npb #tvasahi 0 1 MVP www #allstar 0 1
8 52 8.4: (False Negative) #ALLSTAR 1 0 #npballstar #AllStarGame #AllStar #NPB 1 0 http://t.co/as3ucxvak0 1 0 #Baseball #NPB HR #npb 1 0 #carp #npb 1 0 #allstar 1 0 #allstar 1 0 HR #allstar 1 0 #allstar 1 0 10 #npb 1 0 #allstar 1 0 #allstar 1 0 4 #npb 1 0 2 #AllStar #NPB #npballstar 1 0 #AllStarGame #allstar 1 0 #allstar 1 0 #allstar 1 0 #sbhawks #allstar 1 0 #allstar 1 0 8-3 #allstar 1 0 MVP #allstar 1 0 MVP ( ) #npb 1 0 #allstar 1 0 1 #allstar 1 0 # http://t.co/xqtypiasmn #NPB #baseball 1 0